Prior to this change foreign key constraint references could be left pointing
at tables dropped during operations simulating unsupported table alterations
because of an unexpected failure to disable foreign key constraint checks.
SQLite3 does not allow disabling such checks while in a transaction so they
must be disabled beforehand.
Thanks ezaquarii for the report and Carlton and Tim for the review.
Backport of 315357ad25 from master.
SQLite 3.26 repoints foreign key constraints on table renames even when
foreign_keys pragma is off which breaks every operation that requires
a table rebuild to simulate unsupported ALTER TABLE statements.
The newly introduced legacy_alter_table pragma disables this behavior
and restores the previous schema editor assumptions.
Thanks Florian Apolloner, Christoph Trassl, Chris Lamb for the report and
troubleshooting assistance.
Backport of c8ffdbe514 from master.
Ticket #25619 changed the default protocol to HTTP/1.1 but did not
properly implement keep-alive. As a "fix" keep-alive was disabled in
ticket #28440 to prevent clients from hanging (they expect the server to
send more data if the connection is not closed and there is no content
length set).
The combination of those two fixes resulted in yet another problem:
HTTP/1.1 by default allows a client to assume that keep-alive is
supported unless the server disables it via 'Connection: close' -- see
RFC2616 8.1.2.1 for details on persistent connection negotiation. Now if
the client receives a response from Django without 'Connection: close'
and immediately sends a new request (on the same tcp connection) before
our server closes the tcp connection, it will error out at some point
because the connection does get closed a few milli seconds later.
This patch fixes the mentioned issues by always sending 'Connection:
close' if we cannot determine a content length. The code is inefficient
in the sense that it does not allow for persistent connections when
chunked responses are used, but that should not really cause any
problems (Django does not generate those) and it only affects the
development server anyways.
Refs #25619, #28440.
Regression in ac756f16c5.
Backport of 934acf1126 from master.
sys.stdin.read() blocks waiting for EOF in shell.py which will
likely never come if the user provides input on stdin via the
keyboard before the shell starts. Added check for a tty to
skip reading stdin if it's not present.
This still allows piping of code into the shell (which should
have no TTY and should have an EOF) but also doesn't cause it
to hang if multi-line input is provided.
Backport of 4e78e389b1 from master.
SessionBase.decode() is the inverse operation to SessionBase.encode().
As SessionBase.encode() always returns a string, SessionBase.decode()
should always be passed a string argument. Fixed the file backend, which
was the only backend still passing a bytestring.
Backport of bdae19cf63 from master
The mysqlclient cursor attribute `_last_executed` is always stored as
bytes. Decode it.
TextField values are already type str. No need to decode.
Backport of efd8a82e26 from master
Mariusz Felisiak
Tim Graham
MaximZemskov
Tom Hacohen
Konstantin Alekseev
Simon Charette
Carlton Gibson
Basil Dubyk
takaaki shimbo
Florian Apolloner
Claude Paroz
Mathew Payne
minusf
Sergey Fedoseev
Adam Allred
aspalding
Taoufik
Jon Dufresne
Stephen James
Oleg
Alexander Holmbäck